Coronavirus

[EinSV]

Interesting tidbit from the press conference on the New York report.

According to Governor Cuomo, the higher rates of infection detected statewide for African Americans and Hispanics is primarily due to the fact that they live in NYC where the rates of infection are high, as opposed to differences among races in NYC or upstate:

"This reflects more the regional breakdown. African-Americans and Hispanics in this survey are disproportionately from New York City and New York City is at 21 percent. So the African-American number, the Latino number is at 22 percent and the upstate whites -- they're talking about more upstate -- whites which is 9 [percent] -- but it's 3.6 [percent upstate] in the survey."​

QuickTake by Bloomberg on Twitter

On the flip side, it sounds like the lower rate for whites is being driven by a significantly lower rate of infection upstate (3.6%).

I'm sure there will be more details in the preprint, if not sooner

 

[Daniel in SD]

WTF is wrong with him? This puts the IFR right where we always feared it would be.

I mean, I guess NY is close to herd immunity. Just have to kill off another 30k-40k people in the city, and then we'll be there. Hooray.

I'm mystified how someone so smart could possibly have thought this a month ago.

 

[MXLRplus]

View attachment 535077
I'm mystified how someone so smart could possibly have thought this a month ago.

You are not doing your calculations in Elon Time. This means April 2021. Maybe.

 

[Norbert]

Another number to take into account is the test specificity, if we're trying to modify % positive/bias it by a couple % with these various factors.

The NY State test appears to have 93%-100% specificity. Kind of a wide range. Most likely it is closer to 97%, but I don't see any calibration data.

https://coronavirus.health.ny.gov/s...adsworth-centers-assay-for-sars-cov-2-igg.pdf

So you would take 3% off the infection rate?

 

[azaz]

View attachment 535077
I'm mystified how someone so smart could possibly have thought this a month ago.

he has an incredible amount of hubris

 

[AlanSubie4Life]

View attachment 535077
I'm mystified how someone so smart could possibly have thought this a month ago.

Yeah, really makes you wonder. He clearly sticks to his priors.

So you would take 3% off the infection rate?

I honestly have no idea! These are presumably 95% CI numbers. You'd have to know the sensitivity as well, of course. And then run all the statistics on it.

This is preliminary data, but without the actual details it's hard to say what to do.

I would GUESS that the numbers out of the study would need to be adjusted downwards a couple % before making all the other adjustments discussed. But if there is a sensitivity issue, that would tend to push it back upwards. Etc.

It's really hard. But from my perspective, I just wanted to know whether we were in the range of 0.7 to 1.3% IFR with reasonably high certainty. It certainly seems extremely likely we'll be above that lower number, which makes me sad.

 

[dfwatt]

Along those lines . . .
She got a forgivable loan. Her employees hate her for it.

Needless to say this re-enforces my views against UBI.

At the other end of the extreme you have Florida which has only given out 7% of its benefit to the unemployed that have documented or in many instances serially and futilely attempted to document their unemployment status. Somewhere in the middle is reasonable support.

 

[dfwatt]

he has an incredible amount of hubris

Personally I hope that the explanation is that his account was hacked by either Chinese or Russian trolls or from one of those liberate Virginia Confederacy guys.

You just hate to see somebody who's used to being the smartest guy in the room look like the dumbest guy in the room.

 

[Norbert]

But from my perspective, I just wanted to know whether we were in the range of 0.7 to 1.3% IFR with reasonably high certainty. It certainly seems extremely likely we'll be above that lower number, which makes me sad.

Why sad? It just makes it clear that we can't go for herd immunity. Which we shouldn't anyway. The number of deaths doesn't increase because of that. As far as I can understand.

 

[AlanSubie4Life]

Why sad? It just makes it clear that we can't go for herd immunity.

Well, I would have preferred to see data that suggested an IFR of 0.5% or lower. That would have meant we were closer to herd immunity. Not that we should have gone for herd immunity ever (I never thought we should!), but if we got pushed in that direction or the virus proves impossible to control in this country (due to incompetence), then it would mean fewer deaths.

And I would prefer as low an IFR as possible, since my family, my friends, and I have not yet been infected.

 

[MXLRplus]

Why sad? It just makes it clear that we can't go for herd immunity. Which we shouldn't anyway. The number of deaths doesn't increase because of that. As far as I can understand.

It's not like we get a choice. Either we come up with an effective vaccine, or we risk exposure.
You can shut down the world for 20 years and the virus will not die without human intervention.

 

[Daniel in SD]

So you would take 3% off the infection rate?

Note that upstate New York only had 3.6% positive results. It seems likely that the test is specific enough. Sensitivity is the real question.
They say 3-4 weeks to develop IgG antibodies but obviously it's not a binary thing. The shape of the curve matters a lot in determining when the past infection numbers matched the current antibody numbers.
Another note I haven't seen mentioned here, these antibody tests were taken earlier this week.
I do tend to feel that the lag time between infection and death is on the same order as the lag time between infection and antibodies so your crude calculations may be close to correct. I lot of noise on everything though!

 

[Daniel in SD]

Well, I would have preferred to see data that suggested an IFR of 0.5% or lower. That would have meant we were closer to herd immunity. Not that we should have gone for herd immunity ever (I never thought we should!), but if we got pushed in that direction or the virus proves impossible to control in this country (due to incompetence), then it would mean fewer deaths.

And I would prefer as low an IFR as possible, since my family, my friends, and I have not yet been infected.

Also, the closer you get to herd immunity the more effective mitigation measures are (by lowering the effective reproduction rate below R0).

 

[AlanSubie4Life]

Note that the upstate New York only had 3.6% positive results.

Yeah, best to focus on the places with the very high results for now. Hopefully we'll get more information on the test statistics and larger sample size as time goes by.

 

[Norbert]

Well, I would have preferred to see data that suggested an IFR of 0.5% or lower. That would have meant we were closer to herd immunity. Not that we should have gone for herd immunity ever (I never thought we should!), but if we got pushed in that direction or the virus proves impossible to control in this country (due to incompetence), then it would mean fewer deaths.

We have to stick with the realization that South Korea proved it possible. I never saw an alternative.

And I would prefer as low an IFR as possible, since my family, my friends, and I have not yet been infected.

Same here (except for one young relative with mild symptoms). How does a low IFR help in that situation? I would think a low IFR corresponds to a higher infection rate, which makes it more likely to get infected, but maybe there are different ways of looking at that.

 

[Norbert]

Also, the closer you get to herd immunity the more effective mitigation measures are (by lowering the effective reproduction rate below R0).

However we don't know yet how well and how long immunity is effective.

 

[AlanSubie4Life]

Thread from Trevor on the seroprevalence study of the day:

Trevor Bedford on Twitter

 

[Norbert]

I do tend to feel that the lag time between infection and death is on the same order as the lag time between infection and antibodies so your crude calculations may be close to correct.

Usually I think in terms of lag between test and death, not infection and death. I'll keep my eyes open in that regard.

I lot of noise on everything though!

Yes, I'd rather not get involved in calculating error bars. Over time it will get clearer.

 

[Norbert]

It's not like we get a choice. Either we come up with an effective vaccine, or we risk exposure.
You can shut down the world for 20 years and the virus will not die without human intervention.

Yes, the only choice is to do as South Korea (or better) once we have mitigated down to sane levels. South Korea has a total death number of 240, which translated to the US would be about 1,550, in total at the end of the curve. And they are now down to a very low "maintenance" level ( < 10 new cases per day) which gives them a lot of time to develop a vaccine.

 

[traxila]

Yes, the only choice is to do as South Korea (or better) once we have mitigated down to sane levels. South Korea has a total death number of 240, which translated to the US would be about 1,550, in total at the end of the curve. And they are now down to a very low "maintenance" level ( < 10 new cases per day) which gives them a lot of time to develop a vaccine.

Something we could have done from the start as well, given the appropriate actions and proper mobilizations from our leaders.

Savings: Thousands of lives, trillions of dollars, the economy for the next three years.

Cost. I dunno. A few bill maybe, since some kind of graft is inevitable.